Re: Extremely preliminary WWW benchmark numbers with ISS test3

Rogier Wolff (R.E.Wolff@BitWizard.nl)
Thu, 6 Mar 1997 17:12:18 +0100 (MET)


David S. Miller wrote:
> Average Std Dev Minimum Maximum
>
> Connect time (sec) 0.114010 0.712250 0.000793 42.033459
> Response time (sec) 0.396090 0.825057 0.032083 42.336637
[deletia]
> So, someone who knows something about these things, tell me do those
> numbers suck raw eggs or what?

There is something odd going on. Your server has been "dead" for 42
seconds during this run... Several lost packets (TCP?) on your ethernet?
[grabs stevens] Retransmits should be occurring at 1,3,6,12 and 24 second
intervals after the first packet hits the bucket. That adds up to 46
seconds.

I don't understand.

You should see something like

Connect time (sec) 0.114010 0.0312250 0.000793 0.33459

meaning that the connect time was about 0.1 second, and the longest took
0.3 seconds. Having the longest take 42 seconds is absurd. You've lost
a client when that is your commercial webserver.

A "Maximum" of about 1 second could still be acceptable if you get the
server to swap substantially somewhere during the test. But 42 seconds
is absurd.

Roger.